CBX3 Promotes Colon Cancer Cell Proliferation by CDK6 Kinase- Independent Function During Cell Cycle

www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 12), pp: 19934-19946

Research Paper CBX3 promotes colon cancer cell proliferation by CDK6 kinase- independent function during cell cycle

Yao Fan1,2, Haiping Li1,2, Xiaolong Liang1,2, Zheng Xiang1,2 1Chongqing Key Laboratory of Department of General Surgery, Chongqing Medical University, Chongqing, China 2Department of Gastrointestinal Surgery, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China Correspondence to: Zheng Xiang, email: [email protected] Keywords: colon cancer, CBX3, CDK6, p21, proliferation Received: August 01, 2016 Accepted: December 12, 2016 Published: February 10, 2017

ABSTRACT

Heterochromatin protein 1γ (CBX3) links histone methylation marks to transcriptional silence, DNA repair and RNA splicing, but a role for CBX3 in cancer remains largely unknown. In this study, we show that CBX3 in colon cancer cells promotes the progression of the cell cycle and proliferation in vitro and in vivo. Cell cycle (G1 phase to S phase) related gene CDK6 and p21 were further identified as targets of CBX3. In addition, we found that enhancing CDK6 suppresses cell proliferation by upregulating inhibitor p21 in the absence of CBX3, and this function is independent of the kinase activity of CDK6. Our results demonstrate a key role of CBX3 in colon carcinogenesis via suppressing the expression of CDK6/p21, which may disrupt the role of CDK6 in transcriptionally regulating p21, as part of a negative feedback loop to limit CDK6 excessive activation.

INTRODUCTION Cyclin dependent kinases (CDKs) are core parts of cell cycle regulation, containing over 20 members [16]. The histone proteins as well as several other proteins CDKs in binding and catalyzing cyclins are necessary to are essential elements of heterochromatin formation, and drive progression of the cell cycle. CDK4/CDK6-cyclinD are associated with gene transcriptional repression or complexes enable the cells to pass from G1 to S phase, silencing [1]. The mammalian heterochromatin protein 1 but these proteins are critically required for proliferation is a highly conserved non-histone chromosomal protein only in selected cells types [17, 18]. Experiments with and a major component of heterochromatin. Its orthologs gene-deficient mice for CDK4 or CDK6 are viable [19], include CBX5, CBX3 and CBX1, which are concentrated then lacking CDK4 and CDK6 die of severe anemia and at pericentric heterochromatin or euchromatic sites, hematopoietic abnormalities [20]. Partial experiments each containing a chromodomain [2]. Methylation of demonstrate D-cyclins enhancing interaction with CDK2 histone H3 lysine 9 (H3K9) can be recognized by the or cyclinD-CDK4/CDK6-independent mechanism. CBX3 chromodomain [3]. Dynamic CBX3 binding to However, several are also involved in transcriptional H3K9 is dependent on the histone methyltransferase regulation, DNA damage repair, stem-cell self-renewal, Suv339H1 and maintains stable heterochromatin domains metabolism, spermatogenesis and neuronal function to regulate gene expression [4]. In addition, CBX3 can [21]. Recent researches focus on the role of CDK6 in recruit various cofactors which perform functions in transcriptional regulation. CDK6 acts as a cofactor intracellular biological processes by binding methylated targeting gene promoter, including IL-1, IL-8, EGRI [22– H3K9. These functions include RNA alternative splicing, 25]. The most prominent report that a kinase-independent DNA damage response, transcription elongation, cell function of CDK6 links cell cycle to tumor angiogenesis growth and differentiation [5–10]. Recently, although gene shows CDK6 participating in p16 and VEGF-A expression assays have revealed that CBX3 is linked with transcriptional regulation in lymphoid malignancies [26], lung cancer, osteosarcoma, gastric cardia adenocarcinoma CDK6 acts as a transcriptional regulator on the p16INK4a and colorectal cancer, among other diseases [11–15], the promoter providing an internal safeguard by negative specific mechanisms by which CBX3 promotes cancer feedback loop. However, CDK6 kinase-independent progression are poorly understood. function in colorectal cancer has never been reported.

www.impactjournals.com/oncotarget 19934 Oncotarget To investigate the mechanism by which CBX3 plasmid was constructed as shown in (Supplementary promotes colon cancer, we deleted CBX3 in HCT116 with Figure 1). Several colony cell lines were established by the CRISPR/Cas9 editing system and found that cell cycle from transfected cells. We found that the expression of progression was curbed in G1 phase and proliferation was CBX3 in the cell line, HCT116-B3 was significantly inhibited. Gene expression analysis reveals that CDK6/ lower than that in others (Figure 1A, 1B, 1C and 1D). p21 is upregulated. In addition, CDK6 is able to directly Furthermore, no fluorescence was detected in HCT116-B3 regulate p21 expression in a kinase-independent manner. cells by cell immunofluorescence when immunostained Therefore, we identify a new mechanism by which CBX3 with anti-CBX3 fluorescence antibody (Figure 1E). To promotes colon cancer progression via the CDK6/p21 verify that the deletion of CBX3 was due to CRISPR/cas9 pathway during cell cycle. editing, CBX3 genome was sequenced in HCT116-WT, B1, B3 and B4 cells. We found that an additional base T RESULTS was inserted in the NGG site to change the CBX3 genome sequence (Figure 1F). CRISPR/Cas9 induced CBX3 genome mutation CBX3 promoted colon cancer cell cycle To identify the relationship between CBX3 and the progression and proliferation in vitro pathological process of colon cancer, the CRISPR/cas9 genome editing system was used to modulate the CBX3 To investigate the effect of the deletion of CBX3 gene in HCT116 cells. sgRNA was designed and Px330 on colon cancer cell growth, we used flow cytometry to

Figure 1: CRISPR/cas9 editing system induced the deletion of CBX3 protein. A. Quantitative real-time PCR (qRT-PCR) analysed mRNA level of CBX3 in HCT116-B1, B2, B3, B4 and B5 colony cell lines. The mean±SD were shown from three independent experiment, **P<0.01 compared with control. B. Immunoblot for CBX3 of five colony cell lines derived from transfecting px330 plasmid in which Lmnb1 served as a loading control. C. qRT-PCR analysis mRNA level of CBX3 from HCT116-B3 cell lines continuously cultured for 5 generations. The mean±SD were shown from three independent experiments. **P<0.01 compared with the control. D. Immunoblot for CBX3 of HCT116-B3 and HCT116-WT cells, ACTB served as a loading control. E. Cell immunofluorescence was performed in HCT116-B3 and HCT116-WT cells. Nuclei were identified by DAPI, immunofluorescence staining for CBX3. Scale bar=10 um.F. Base T was inserted in NGG site in B1, B3 and B4, base G mutation is base A in B4, verified by sequencing the CBX3 genome. www.impactjournals.com/oncotarget 19935 Oncotarget analyze cell cycle and apoptosis. These results shown after incubation for 5 days (Figure 2C and Supplementary that cell cycle was significantly inhibited in G1 phase Figure 3B). These results were further confirmed by compared with the control (Figure 2A and Supplementary EdU staining which could detect nucleotide analogue Figure 3A), but we did not find any change in apoptosis incorporation into replicated DNA (Figure 2D). These (Supplementary Figure 2A and 2B). Cell proliferation was data indicate that CBX3 promotes colon cancer cell also analyzed by MTS assay, and the deletion of CBX3 proliferation by curbing cell cycle G1-S phase transition. in colon cancer cells evidently reduced cell growth rate (Figure 2B and Supplementary Figure 3C). To directly The deletion of CBX3 directly enforces the observe cell growth, colony formation assays were expression of CDK6 and p21 performed. A specified number of cells were seeded into plates and the deletion of CBX3 in the colon cancer cell To investigate the underlying mechanism, we line significantly reduced the numbers of colonies per dish examined the expression of several genes known to be

Figure 2: CBX3 deletion inhibited cell cycle progression and proliferation in vitro. A. Flow cytometry assays were performed to analyze cell cycle in HCT116-WT and B3 cells. Values at different stages of cell cycle represent mean±SD from three independent experiments. **P<0.01 compared with control. B. Colon cancer cell proliferation were detected by MTS and absorbance at 490 nm at different time points is shown. Each data point represents mean±SD from three independent experiments. **P<0.01 compared with control. C. Colony formation assays were performed using HCT116-WT and HCT116-B3 cells. Colony number were analyzed by measuring absorbance at 570 nm, The results were shown as mean±SD from three independent experiments. *P<0.05 compared with control. D. The effect of CBX3 deletion on the growth of colon cancer HCT116 was analyzed by EdU proliferation assay. **P<0.01 compared with control. www.impactjournals.com/oncotarget 19936 Oncotarget important for cell cycle G1-S phase transition, including in tumor growth between HCT116-B3 and HCT116-WT CDK2, CDK4, CDK6, p16, p21, CDH3, CDC25A, (Figure 4A, 4B, 4C and 4D). These results suggest that CCND1, CCNE1, E2F1 and E2F3 in HCT116-B3 and CBX3 accelerates tumor formation in vivo. HCT116-WT cells (Figure 3A). We found the mRNA We had identified that the deletion of CBX3 abundances of CDK6 and p21 consistently increased increases CDK6 and p21 expression in colon cancer in the deletion of CBX3 cells (Figure 3B). Immunoblot cells. To further confirm this result, immunoblots were results also revealed that the deletion of CBX3 causes performed for CBX3, CDK6 and p21 in xenograft tumor high protein expression of CDK6 and p21 compared tissue. We found that deletion of CBX3 significantly with control (Figure 3C). In order to determine whether increased CDK6 and p21 protein abundances (Figure elevated CDK6 and p21 in the cells in which CBX3 was 4E). In addition, these results were further confirmed by deleted could be observed directly at the level of a single immunohistochemistry (Figure 4F and 4G). These data cell. Cell immunofluorescence staining for CDK6 and p21 indicate that CBX3 promotes colon cancer progression showed a dramatic difference between HCT116-B3 and correlation with CDK6 and p21 expression. HCT116-WT (Figure 3D). Decreased CDK6 expression promotes cell CBX3 promotes xenograft tumor growth in proliferation in the absence of CBX3 correlation with CDK6 and p21 expression Although CDK6 and p21 are important regulators To investigate the mechanism by which CBX3 of the G1-S cell cycle transition, they play opposing directly promotes colon cancer cell proliferation in vivo, roles in cell cycle regulation. The study conducted by we performed xenograft tumor growth assays in nude mice Kollmann et al. reveals an unexpected role for CDK6, and the size and weight of tumors were measured after as a transcriptional regulator on the p16ink4a promoter 24 days. We found that there is a significant difference in BCR-ABL transformed B-cell leukemia/lymphoma

Figure 3: Deletion of CBX3 reduced the expression of CDK6 and p21. A. The mRNA abundances of CDK2, CDK4, CDK6, P16, P21, CDH3, CDC25A, CCND1, CCNE1, E2F1, E2F3 were analyzed with qRT-PCR in HCT116-B3 cells relative to the control. The results were shown as the mean±SD from three independent experiments. **P<0.01 compared with control. B. qRT-PCR were performed to analyze the mRNA level of CDK6 and p21 of HCT116-B3 relative to control. The results were shown as the mean±SD from three independent experiments. **P<0.01 compared with control. C. Immunoblot for CDK6, p21 and CBX3 derived from HCT116-WT, B1, B3 and B4 cells,, with GAPDH served as a loading control. D. Cell immunofluorescence were performed using HCT116-WT and B3 cells. Nuclei were identified by DAPI, immunofluorescence staining for CDK6 and p21. Scale bar=10 um. www.impactjournals.com/oncotarget 19937 Oncotarget cells. To investigate the specific biological role of CDK6 CDK6 regulates the expression of p21 by the in colon cancer, we knocked down CDK6 expression in kinase-independent function HCT116 and HCT116-B3 cells. We found that cell cycle G1-S phase transition was promoted in the absence of We had previously found that CBX3 can inhibit CBX3, but cell cycle was curbed in G1 phase in HCT116 the expression of CDK6 and P21 in colon cancer. To (Figure 5A). In addition, cell proliferation was analyzed investigate the underlying mechanism, we examined p21 by MTS assay and colony formation. These results shown mRNA and protein abundances after increasing CDK6 that cell growth in HCT116-B3-shCDK6 is faster than in expression. We found that enhanced CDK6 expression was control, but cell growth in HCT116-shCDK6 is slower consistently accompanied by elevated tumor suppressor than in control (Figure 5B and 5C). Taken together, our p21 in the absence of CBX3 (Figure 6A and 6B) and the data indicates that CDK6 knocked down promotes colon expression level of p21 was significantly reduced when cancer cell proliferation in the absence of CBX3. CDK6 was knocked down (Figure 6C and 6D), but we

Figure 4: CBX3 deletion suppressed xenograft tumors growth in nude mice. A. The growth of subcutaneous tumors in nude mice from HCT116-B3 or control cell growth (n=10). B. Tumor tissues derived from xenograft tumors in nude mice after 24 days. Top, tumor from control; bottom, tumor from the deletion of CBX3. C. The mean volume of xenograft tumors from HCT116-B3 versus WT, **P<0.01. D. The average tumor weight from HCT116-B3 versus WT, **P<0.01. E. xenograft tumors tissues protein were extracted from HCT116-B3 (n=4,#A,#B,#C,#D) and HCT116-WT (n=4,#E,#F,#G,#H), Immunoblot for CDK6, P21 and CBX3. GAPDH served as a loading control. F. Immunohistochemical stainings of xenograft tumor tissues from HCT116-B3 (#D) and control HCT116-WT (#E) for CDK6 and P21. Original magnification×400. G. Quantification of CDK6 or p21 positive cells from F picture, **P<0.01 compared with control. www.impactjournals.com/oncotarget 19938 Oncotarget did not find any change in HCT116. Meanwhile, CDK6 of cell proliferation correlating with increasing inhibitor induced the expression of VEGF had not been observed concentration (Figure 6E). However, inhibitor treatment (Supplementary Figure 3D). did not affect the abundance of p21 mRNA (Figure 6F). Due to CDK6 and p21 performing opposing The results suggest that CDK6 regulates the expression of functions in cell cycle progression, we speculated p21 by the kinase-independent activity. that CDK6 may not act as a kinase dependent function To study the molecular mechanism of CDK6 cooperating with p21 to affect cell proliferation, but it regulation of p21 expression, a luciferase reporter assay was the kinase-independent function. A small molecule was performed to analyze whether CDK6 enhanced inhibitor (PD0332991) had been identified to directly the p21 promoter activity. The reporter plasmid was bind the ATP binding pocket domain of CDK4 and constructed and respectively cotransfected with two CDK6. When we treated cells in HCT116-B3 with protein expression plasmids, including pCDNA3.1- CDK4/CDK6 inhibitor, we found a significant reduction CDK6 and pCDNA3.1-CBX3. We found that CDK6

Figure 5: Decreased CDK6 expression promoted cell proliferation in the absence of CBX3. A. Cell cycle was analyzed by flow cytometry from CDK6 knocked down in HCT116 and HCT116-B3. Values were shown at different stages of cell cycle representing mean±SD from three independent experiments. *P<0.05 compared with control. B. Colony formation assay of HCT116 and HCT116-B3 with CDK6 knocked down or scrambled control cells. *P<0.05. C. CDK6 knocked down or scrambled control cell proliferation were detected by MTS. Absorbance of 490 nm were shown at different time points represented as mean±SD from three independent experiments. **P<0.01 compared with control. www.impactjournals.com/oncotarget 19939 Oncotarget could significantly increase p21 promoter activity. by reduced or undetectable expression of p21, where However, enhanced CBX3 expression markedly reduced CBX3 expression was significantly upregulated in colon p21 promoter activity (Figure 6G). Our data indicate that cancer tissue (Figure 7A, 7B, 7C, 7D). However, we CBX3 prevents transcriptional regulation of CDK6. observed the opposite results in non-cancerous tissues. These results indicate that inverse protein expression Inverse protein expression trends of CDK6 and levels of CDK6 and p21 may be regulated by CBX3 to p21 in human colon cancer tissue disrupt a negative feedback loop in colon cancer.

To further confirm that CBX3 promotes colon cancer DISCUSSION progression via regulating the expression of CDK6/p21, we examined the expression of CBX3, CDK6 and p21 in CBX3, from the heterochromatin protein 1 family, human non-cancerous tissue (NCA) and cancerous tissue whose members are closely related to the biological (CA). We found high abundances of CDK6 accompanied processes in cells and have been proposed to be driving

Figure 6: CDK6 regulated p21 expression in the absence of CBX3. A. qRT-PCR analysed CDK6 and p21 mRNA abundances in HCT116 and HCT116-B3 with CDK6 overexpression, the mean±SD from three independent experiments were shown. **P<0.01 compared with control. B. Immunoblot for CDK6 and p21 of HCT116 and HCT116-B3 with CDK6 overexpression. C. CDK6 and p21 mRNA level of HCT116-B3 cells with CDK6 knocked down were analyzed with qRT-PCR, the mean±SD from three independent experiments are shown.**P<0.01 compared with control. D. Immunoblot for CDK6 and p21 of HCT116-shCDK6, B3-shCDK6 and control cells. E. Dose- response curve of HCT116-B3 cells treated for 24 hr with the CDK6/4 inhibitor PD0332991 (n = 3). F. p21 mRNA levels of HCT116-B3 cells treated for 24 hr with 0, 30, 100, 300, and 1000 nM PD0332991 were analyzed by qPCR. P>0.05 compared with control. G. Luciferase reporter assays were performed in HCT116-B3 cells, co-transfected with pGL3-basic-p21 and pCDNA3.1-CDK6 or pCDNA3.1-CBX3, Empty pGL-basic and PCDNA3.1 plasmid was used as a negative control. *P<0.05. www.impactjournals.com/oncotarget 19940 Oncotarget forces for some diseases when they are frequently where CDK6 and p21 are consistently upregulated. expressed at aberrant levels. Although methylation of On first thought, the result might appear paradoxical, H3K9 is considered to be a characteristic of CBX3 because for a long time CDK6 was regarded as a mere elevation, the underlying mechanism for the preferential homolog of CDK4 with overlapping functions in terms upregulation of CBX3 in carcinogenesis remains a of Rb phosphorylation. The G1 kinases CDK4 and mystery. In this study, we show that CBX3 protein levels CDK6 are true cell-cycle kinases that regulate exit are significantly increased in primary colon cancer tissues. from the G1 phase of the cell cycle. They have received Moreover, the higher expression of CBX3 is detected in considerable attention as major drivers of cancer [21]. poorly differentiated colorectal cancer tissues and the The study conducted by Kollmann et al. reveals several overall patient survival rate was significantly decreased novel roles for CDK6 acting as the kinase-independent in [17]. These results suggest that CBX3 plays a distinctive transcription. Induction of p16 by CDK6 serves as a fail- role in the pathological process of cancer. safe mechanism that restricts excessive activity of CDK6 The deletion of CBX3 in colon cancer cells curb by forming a negative feedback loop [34]. In addition, cell cycle progression (G1 phase to S phase) suppress CDK6 can bind to chromatin and recruit binding motifs cell proliferation in vitro and tumor growth in vivo, of transcription factors such as c-Jun, AP-1, P65 to induce

Figure 7: Inverse relationship between CDK6 and p21 expression in human colon cancer. A. Hematoxylin and eosin (H&E) staining and immunohistochemical stainings of CBX3, CDK6 and p21 in colon cancer tissue and non-cancerous tissue. Representative micrographs including colon cancer tissue and non-cancerous tissue were shown, original magnification (×400). B. total IHC score of CBX3, CDK6 and p21 in colon cancer tissue and non-cancerous tissue (n=30), **P<0.01 or *P<0.05 compared with control. C. Immunoblot for CBX3, CDK6 and p21 of human colon cancer tissue (CA) and non-cancerous tissue (NCA) (n=24), GAPDH served as a loading control. D. CBX3, CDK6 and p21 protein expression levels by quantitation of density of protein bands from immunoblot in (C) in CA relative to NCA (n=24, **P<0.01). www.impactjournals.com/oncotarget 19941 Oncotarget the expression of the genes VEGF-A, EGR1, IL. However, CBX3 is in high abundance. Our data indicate that CBX3 cell cycle-independent functions for CDK6 are not shared participates in the silencing of p21 and prevents CDK6 with CDK4 [21]. In our study, we found CDK6 regulates transcriptional regulation. the expression of p21 by promoting its promoter activity Cell growth is strictly regulated by cell cycle when CBX3 is deleted and its expression is not affected progression. CDK4/6–cyclinD complexes are controlled by CDK6 kinase activity. Therefore, our finding can be by two classes of inhibitors. CDK4 and CDK6 single rationalized by postulating a feedback loop including knockout mice are viable. For one thing partially the upregulation of the tumor suppressor and cell-cycle overlapping and redundant physiologic roles for CDK4 inhibitor p21. and CDK6, for another CDK6 may have a unique role CBX3 is an evolutionarily conserved chromosomal to cell in a state, this is particularly evident in stem protein that binds methylated H3K9, where it recruits cells, where CDK6 is not required under homeostatic cofactors or directly interacts with targets [17]. CBX3 is conditions but becomes critical under conditions of stress, also a major component of heterochromatin and maintains such as oncogenic stress [35, 36]. CDK6 function as a the transcriptionally repressive heterochromatin structure transcriptional regulator linking cell-cycle progression to to regulate gene transcription [3], including transcription differentiation and other cellular functions is starting to silencing and retroviral silencing [30–32]. CBX3 has be understood. The study of Kollmann et al. shows that recently been proposed to have a close relationship with CDK6 is capable of two opposing functions: it is able the tumor, such as non-small cell lung cancer, prostate to inhibit or accelerate cell proliferation depending on cancer and breast cancer among others. However, the whether p16 is present or not, but it is unclear what causes underlying mechanism of CBX3 to promote tumorigenesis the initial upregulation of CDK6 and the silencing of p16. is still largely unknown. In pancreatic cancer, NFATc2 We find that p21 is nearly absent from colon cancer tissues, binding to p15 promoter recruits histone methyltransferase but CBX3 significantly increases. Promoter activity of Suv39H1 resulting in H3K9 trimethylation, which allows p21 also significantly decreases with enhanced CBX3 docking of CBX3 to the repressor complex giving rise expression. In addition, the histone marks H3K9me2/3 to p15 silencing [33]. In the absence of CBX3, enforced on the p21 promoter were also significantly reduced CDK6 levels inhibit cell growth by upregulating inhibitor when CBX3 was downregulated [17]. We speculate that p21, but we can't do it in wild type cells, and promoter methylated H3K9 binding CBX3 in the p21 promoter activity is suppressed by CBX3. In addition, we found prevents the function of CDK6 as a transcriptional that histologic sections of human colon malignancies have regulator, in which it silences the expression of p21 and an inverse correlation of CDK6 and p21 expression, but disrupts a negative feedback loop (Figure 8).

Figure 8: A hypothetical schematic sketch of CBX3/CDK6/p21 controlling cell cycle and proliferation. www.impactjournals.com/oncotarget 19942 Oncotarget Table 1: Amplify CDK6 primer and real-time PCR primer CDK6 F: GGAATTC ATGGAGAAGGACGGCCTGTGCC (ECOR1) R: CGGGATCC TCAGGCTGTATTCAGCTC (Bamh1) name primer name primer F:TGGCCTCCAACAAAACTACA F:CCAGGCAGGCTTTTCATTCA CBX3 CDK6 R:TCCCATTCACTACACGTCGA R:AGGTCCTGGAAGTATGGGTG F: ACAGATCATCCGCAAACACG F: GTATGAGCTCTTTGGCCACG CCND1 CDH3 R:GGCGGTAGTAGGACAGGAAG R: ATGGCATCATCCTCATCCGT F:GGAAGAGGAAGGCAAACGTG F: CTACTGATGGCAAGCGTGTC CCNE1 CDC25A R: TTTGTCAGGTGTGGGGATCA R: TCTCTCTCACATACCGGCAC F: TGTCTTGTACCCTTGTGCCT F:CTTCGTAGCATTGCAGACCC P21 E2F1 R: GGCGTTTGGAGTGGTAGAAA R:AAAACATCGATCGGGCCTTG F: TCCGGATCTTTCGGACTCTG F: ATCCCTAAACCCGCTTCCAA CDK2 E2F3 R: ACAAGCTCCGTCCATCTTCA R:GAGGCCAGAGGAGAAAGGTT F: CTTCCCATCAGCACAGTTCG F: CGATGTCGCACGGTACCT CDK4 P16 R: GGGGTGCCTTGTCCAGATAT R: GACCTTCCGCGGCATCTAT

In conclusion, these findings provide a new weeks were purchased from Animal Experimental Center perspective for understanding CBX3 as an oncogenic of Chongqing medical university. All animal care and molecule in colon cancer. We have identified a novel handling procedures were handled in accordance with role for CDK6 as a transcriptional regulator to link cell- study protocols approved by the Ethics Committee of cycle progression in colon cancer. Although we have Chongqing Medical University. demonstrated that CDK6 and p21 are regarded as the target of CBX3 regulation, it remains to be identified Construction of CRISPR/Cas9 vector whether other inhibitors also are related to the CBX3 regulated network. The CBX3 genome sequence was found in NCBI web-based database. The sgRNA was designed MATERIALS AND METHODS using the SAM Cas9 activator design tool (http:// sam.genome-engineering.org/database/). Two sgRNA Materials and reagents oligos were synthesized by Invitrogen. We followed the SAM target sgRNA cloning protocol to ligate the guide Colon cancer cells were obtained from the tumor target sequence into pX330-U6-Chimeric_BB-CBh- laboratory of the First Affiliated Hospital of Chongqing hSpCas9 vector [27]. Px330 plasmids are sequenced Medical University, China. The plasmid pX330-U6- using primers: F: AGGGATGGTTGGTTGGTGGG, R: Chimeric_BB-CBh-hSpCas9 was a gift from Feng Zhang CCAATCCTCCCCCTTGCTGT. (Addgene plasmid # 42230). Lentivirus plasmid pLKO.1 puro was a gift from Bob Weinberg (Addgene plasmid # Cell culture, transfections, cell clone, and 8453). Rabbit anti-CBX3 (A2248) was purchased from infection ABclonal Biotech Co (ABclonal, USA). The following antibodies were purchased from Abcam (Cambridge, HCT116 cell lines were purchased from ATCC MA): rabbit anti-CDK6 (ab124821), rabbit anti-p21 and cultured in DMEN (Gibco, USA) with 10% fetal (ab109520). Rabbit anti-GAPDH (10494-1-AP) and bovine serum (BI, ISR) and incubated at 37°C in a rabbit anti-β-tubulin (10094-1-AP) were purchased from humidified atmosphere containing 5% carbon dioxide. Proteintech Group, Inc. (Proteintech, China). Colon Each of the plasmids was transfected into the cell line cancer tissue samples and matched adjacent normal with Lipofectamine 2000 (Invitrogen) according to tissues were derived from patients undergoing surgical the manufacturer's instructions. Px330 plasmid was procedures at the First Affiliated Hospital of Chongqing transfected into HCT116. After 48 hours, the cells were Medical University, China. All patients in the study cultured in a low density by trypsin digestion cell and add provided written consent, and were approved by the the puromycin to screen cells. Clone cell lines were picked Ethics Committee from the First Affiliated Hospital of using 10ul pipette tips into 96 well plates where cells Chongqing Medical University. Nude mice ages 4 to 5 are continuously cultured and gradually expanded. We www.impactjournals.com/oncotarget 19943 Oncotarget obtained one cell line which we numbered HCT116-B3. Tumor xenografts in nude mice Human CDK6 editing sequences were amplified by PCR 6 with the following primers (see Supplemental Material, 3×10 cells were diluted in 100 ul PBS, and each Table 1), which were cloned into the lentiviral vector mouse was injected subcutaneously on the left side of the plasmid pCDH-CMV-MCS-EF1-copGFP at unique EcoRI neck. Tumor growth rate was monitored by measuring and BamHI sites. A shRNA for human CDK6 targeting tumor diameter (length=L, width=W) every 4 days. The site was: 5’- GAGTAGTGCATCGCGATCTAA -3’ and tumor volume was calculated using the relationship: 2 was inserted into the SpeI/BamHI sites in the pLKO.1 puro 1/2LW . Mice were handled after 24 days, and tumors lentiviral vector [28]. Lentiviral particles were produced were collected following weight analysis. in HEK293T cells and infected HCT116 and HCT116-B3 cells. Immunohistochemical staining

Protein extraction and immunoblotting All tissues were fixed with 10% paraformaldehyde, and embedded in paraffin wax. Paraffin sections were All of protein extraction use RIPA cell lysis buffer placed in incubators kept at 55°C for 4 hours. The sections (p0013B) from Beyotime Biotechnology according to were immersed in two consecutive washings in xylol for the manufacturer's instructions. Protein samples (30- 20 min to remove paraffin. Sections were then hydrated 50 μg) were separated by SDS-PAGE, transferred onto with different concentrations of ethanol including 100%, PVDF membranes, and blocked with 5% milk. Primary 95%, 85%, 70% and deionized water respectively. The antibodies were diluted with 5% BSA and 4°C overnight. sections were immersed in citrate buffer solution (0.01 Following incubation, the appropriate secondary mol/L, pH 6.0) and heated to repair antigen, then 0.5% antibodies were used from Proteintech (SA00001-2). Triton-x-100 was incubated 30 min after washing in Finally, chemiluminescence was visualized with Western PBS. Biotin-streptavidin HRP detection systems (ZSGB, Bright ECL HRP substrate (Advansta, USA). Equal China) were then used to stain the section according to the loading of protein was verified by immunoblotting with manufacturer's instructions. In parallel, tissue samples in anti-β-tubulin or anti-GAPDH antibody. which the primary antibody was replaced by PBS served as negative control. RNA extraction, cDNA synthesis, and real-time Immunofluorescence PCR Cells were incubated on the glass coverslip. Total RNA was isolated with Trizol reagent Cells were fixed with fresh 4% paraformaldehyde, (Invitrogen, USA) according to the manufacturer's permeabilized and blocked with 0.5% Triton-X-100 and instructions. cDNA was synthesized from isolated 4% BSA, respectively. Primary antibody and secondary RNA using PrimeScript RT Reagent Kit with gDNA antibody were diluted with 4% BSA. DNA was stained Eraser (TAKARA, Japan) following the manufacturer's with 4′,6-diamidino -2-phenylindole (DAPI) (Beyotime, instructions. Real-time PCR was performed with China). A fluorescence microscope was used to detect the UtraSYBR Mixture (CWBIO, China), and the relative targeted protein. levels of gene expression were determined by comparison with β-actin expression. The primer sequences used for qRT-PCR are listed in Table 1. Luciferase assay The pGL3-Basic, pCDNA3.1and pRL-TK Cell cycle analysis, MTS and EDU cell plasmids were purchased from Invitrogen (USA). proliferation assay, colony formation assay The p21 promoter sequence was identified in the NCBI web database and the sequence (-1200- Cell cycle analysis was performed by flow +300) was amplified using the following primers: cytometry from Academy of Life Sciences (Chongqing FCGGCTAGCGACAATGCTTAGTTCAGATAC, R: CC Medical University, China). Cell proliferation was CAAGCTTTACCCAGACACACTCTAAGG. HCT116 quantified by CellTiter 96® Aqueous One Solution Cell cells were co-transfected with pGL3-basic-P21 and Proliferation Assay (Promega, USA) and incorporation pCDNA3.1-CDK6 or pCDNA3.1-CBX3. Each sample was of 5-ethynyl-20-deoxyuridine (EdU) using an EdU Cell also co-transfected with pRL-TK. Cells were harvested 48 Proliferation Assay Kit (Ribobio, China) according to hr later and assayed with the Dual Luciferase Reporter the manufacturer's instructions, respectively. Colony Assay System (Promega, Madison, WI) according to the formation was observed with a crystal violet cell colony manufacturer's instructions. Relative luciferase activity staining kit and cell growth was quantified by detecting was normalized to renilla luciferase activity. The assay was absorbance value at 570 nm (GenMed, USA). repeated three times in independent experiments.

www.impactjournals.com/oncotarget 19944 Oncotarget Statistical analysis transcription elongation through mammalian chromatin. Mol Cell. 2005; 19: 381-391. All statistical analysis was performed using 9. Morikawa K, Ikeda N, Hisatome I, Shirayoshi Y. SPSS16.0 (SPSS Inc. Chicago, IL, USA). Data were Heterochromatin protein 1gamma overexpression in presented as mean ± SD differences between multiple P19 embryonal carcinoma cells elicits spontaneous means were evaluated by two-tailed Student's t-test. A differentiation into the three germ layers. Biochem Biophys value of P<0.05 was considered statistically significant. Res Commun. 2013; 431: 225-231. IHC-score was performed as described previously [29]. 10. Brown JP, Bullwinkel J, Baron-Luhr B, Billur M, Schneider Student's t-test was performed for IHC-score and a value P, Winking H, Singh PB. Correction: HP1gamma function is of P<0.05 was considered statistically significant. required for male germ cell survival and spermatogenesis. Epigenetics Chromatin. 2012; 5: 18. ACKNOWLEDGMENTS 11. Han SS, Kim WJ, Hong Y, Hong SH, Lee SJ, Ryu DR, Lee W, Cho YH, Lee S, Ryu YJ, Won JY, Rhee H, Park JH, et al. This work was supported by the major project RNA sequencing identifies novel markers of non-small cell of Chongqing Science & Technology Commission lung cancer. Lung Cancer. 2014; 84: 229-235. (cstc2012ggyyjs10044), the Scientist Culture Plan of Yuzhong District of Chongqing (20130120), Wu Jieping 12. Saini V, Hose CD, Monks A, Nagashima K, Han B, fund (320.6750.12685). Newton DL, Millione A, Shah J, Hollingshead MG, Hite KM, Burkett MW, Delosh RM, Silvers TE, et al. Identification of CBX3 and ABCA5 as putative biomarkers CONFLICTS OF INTEREST for tumor stem cells in osteosarcoma. PLoS One. 2012; 7: e41401. The authors declare no conflicts of interest. 13. Thorne JL, Ouboussad L, Lefevre PF. Heterochromatin protein 1 gamma and IkappaB kinase alpha interdependence REFERENCES during tumour necrosis factor gene transcription elongation in activated macrophages. Nucleic Acids Res. 2012; 40: 1. Grewal SI, Elgin SC. Heterochromatin: new possibilities 7676-7689. for the inheritance of structure. Curr Opin Genet Dev. 2002; 14. Xu X, Xu L, Gao F, Wang J, Ye J, Zhou M, Zhu Y, Tao L. 12: 178-187. Identification of a novel gene fusion (BMX-ARHGAP) in 2. Minc E, Allory Y, Courvalin JC, Buendia B. gastric cardia adenocarcinoma. Diagn Pathol. 2014; 9: 218. Immunolocalization of HP1 proteins in metaphasic 15. Liu M, Huang F, Zhang D, Ju J, Wu XB, Wang Y, Wang mammalian chromosomes. Methods Cell Sci. 2001; 23: Y, Wu Y, Nie M, Li Z, Ma C, Chen X, Zhou JY, et al. 171-174. Heterochromatin protein HP1gamma promotes colorectal 3. Bannister AJ, Zegerman P, Partridge JF, Miska EA, Thomas cancer progression and is regulated by miR-30a. Cancer JO, Allshire RC, Kouzarides T. Selective recognition of Res. 2015; 75: 4593-4604. methylated lysine 9 on histone H3 by the HP1 chromo 16. Malumbres M, Harlow E, Hunt T, Hunter T, Lahti JM, domain. Nature. 2001; 410: 120-124. Manning G, Morgan DO, Tsai LH, Wolgemuth DJ. Cyclin- 4. Cheutin T, McNairn AJ, Jenuwein T, Gilbert DM, Singh PB, dependent kinases:a family portrait. Nat Cell Biol. 2009; Misteli T. Maintenance of stable heterochromatin domains 11: 1275-1276. by dynamic HP1 binding. Science. 2003; 299: 721-725. 17. Braden WA, McClendon AK, Knudsen ES. Cyclin- 5. Salton M, Voss TC, Misteli T. Identification by high- dependent kinase 4/6 activity is a critical determinant of throughput imaging of the histone methyltransferase pre-replication complex assembly. Oncogene. 2008; 27: EHMT2 as an epigenetic regulator of VEGFA alternative 7083-7093. splicing. Nucleic Acids Res. 2014; 42: 13662-13673. 18. Kozar K, Sicinski P. Cell cycle progression without cyclin 6. Saint-Andre V, Batsche E, Rachez C, Muchardt C. Histone D-CDK4 and cyclin D-CDK6 complexes. Cell Cycle. 2005; H3 lysine 9 trimethylation and HP1gamma favor inclusion 4: 388-391. of alternative exons. Nat Struct Mol Biol. 2011; 18: 19. Malumbres M, Sotillo R, Santamarfa D, Galan J, Cerezo 337-344. A, Ortega S, Dubus P, Barbacid M. Mammalian cells cycle 7. Akaike Y, Kuwano Y, Nishida K, Kurokawa K, Kajita K, without the D-type cyclin- dependent kinases Cdk4 and Kano S, Masuda K, Rokutan K. Homeodomain interacting Cdk6. Cell. 2004; 118: 493–504. protein kinase 2 regulates DNA damage response through 20. Kozar K, and Sicinski P. Cell cycle progression without interacting with heterochromatin protein 1gamma. cyclinD-CDK4 and cyclin D-CDK6 complexes. Cell Cycle. Oncogene, 2015; 34: 3463-3473. 2005; 4:388–391 8. Vakoc CR, Mandat SA, Olenchock BA, Blobel GA. Histone 21. Tigan AS, Bellutti F, Kollmann K, Tebb G, Sexl V. CDK6-a H3 lysine 9 methylation and HP1gamma are associated with review of the past and a glimpse into the future: from www.impactjournals.com/oncotarget 19945 Oncotarget cell-cycle control to transcriptional regulation. Oncogene. 29. Ma LJ, Wu WJ, Wang YH, Wu TF, Liang PI, Chang 2015; 35: 3083-3091. IW, He HL, Li CF. SPOCK1 Overexpression Confers a 22. Handschick K, Beuerlein K, Jurida L, Bartkuhn M, Müller Poor Prognosis in Urothelial Carcinoma. J Cancer. 2016; H, Soelch J, Weber A, Dittrich-Breiholz O, Schneider H, 7: 467-476. Scharfe M, Jarek M, Stellzig J, Schmitz ML, et al. Cyclin- 30. Maksakova IA, Goyal P, Bullwinkel J, Brown JP, Bilenky dependent kinase 6 is a chromatin-bound cofactor for M, Mager DL, Singh PB, Lorincz MC. H3K9me3 binding NF-kappaB- dependent gene expression. Mol Cell. 2014; proteins are dispensable for SETDB1/H3K9me3-dependent 53: 193-208. retroviral silencing. Epigenetics Chromatin. 2011; 4: 12. 23. Buss H, Handschick K, Jurrmann N, Pekkonen P, Beuerlein 31. Ellis J, Hotta A, Rastegar M. Retrovirus silencing by an K, Müller H, Wait R, Saklatvala J, Ojala PM, Schmitz epigenetic TRIM. Cell. 2007; 131: 13-14. ML, Naumann M, Kracht M. Cyclin-dependent kinase 32. Du Chene I, Basyuk E, Lin YL, Triboulet R, Knezevich A, 6 phosphorylates NF-kappaB P65 at serine 536 and Chable-Bessia C, Mettling C, Baillat V, Reynes J, Corbeau contributes to the regulation of inflammatory gene P, Bertrand E, Marcello A, Emiliani S, et al. Suv39H1 and expression. PLoS One. 2012; 7: e51847. HP1gamma are responsible for chromatin-mediated HIV-1 24. Guzman ML. CDK6 is a regulator of stem cells "Egr" to transcriptional silencing and post-integration latency. Embo wake up. Blood. 2015; 125: 7-9. j. 2007; 26: 424-435. 25. Scheicher R, Hoelbl-Kovacic A, Bellutti F, Tigan AS, 33. Baumgart S, Glesel E, Singh G, Chen NM, Reutlinger Prchal-Murphy M, Heller G, Schneckenleithner C, Salazar- K, Zhang J, Billadeau DD, Fernandez-Zapico ME, Gress Roa M, Zöchbauer-Müller S, Zuber J, Malumbres TM, Singh SK, Ellenrieder V. Restricted heterochromatin M, Kollmann K, Sexl V. CDK6 as a key regulator of formation links NFATc2 repressor activity with growth hematopoietic and leukemic stem cell activation. Blood. promotion in pancreatic cancer. Gastroenterology. 2012; 2015; 125: 90-101. 142: 388-398.e381-387. 26. Kollmann K, Heller G, Schneckenleithner C, Warsch W, 34. Otto T, Sicinski P. The kinase-independent, second life of Scheicher R, Ott RG, Schäfer M, Fajmann S, Schlederer CDK6 in transcription. Cancer Cell. 2013; 24: 141-143. M, Schiefer AI, Reichart U, Mayerhofer M, Hoeller C, et al. 35. Laurenti E, Frelin C, Xie S, Ferrari R, Dunant CF, Zandi A kinase-independent function of CDK6 links the cell cycle S, Neumann A, Plumb I, Doulatov S, Chen J, April C, Fan to tumor angiogenesis. Cancer Cell. 2013; 24: 167-181. JB, Iscove N, et al. CDK6 levels regulate quiescence exit 27. Cong L, Ran FA, Cox D, Lin S, Barretto R, Habib N, Hsu in human hematopoietic stem cells. Cell Stem Cell. 2015; PD, Wu X, Jiang W, Marraffini LA, Zhang F. Multiplex 16: 302–313. genome engineering using CRISPR/Cas systems. Science. 36. Scheicher R, Hoelbl-Kovacic A, Bellutti F, Tigan A-S, 2013; 339: 819-823. Prchal-Murphy M, Heller G, Schneckenleithner C, Salazar- 28. Stewart SA, Dykxhoorn DM, Palliser D, Mizuno H, Yu Roa M, Zöchbauer-Müller S, Zuber J, Malumbres EY, An DS, Sabatini DM, Chen IS, Hahn WC, Sharp M, Kollmann K, Sexl V. CDK6 as a key regulator of PA, Weinberg RA, Novina CD. Lentivirus-delivered stable hematopoietic and leukemic stem cell activation. Blood. gene silencing by RNAi in primary cells. Rna. 2003; 9: 2015; 125: 90–101. 493-501.